CID administration preferentially expands genetically modified erythrocytes. Ten CBA-Pk-1^slc/Pk-1^slc mice were conditioned with 400 cGy radiation and then infused with 5 million F36Vmpl^GFP-transduced CBA/N marrow cells. The mice were observed for 22 weeks, at which time they were paired based on RBC number and percentage of reticulocytes. The cohort with lower GFP expression (8.6% ± 1.4% vs 17.6% ± 8.1%) was treated with AP20187, 10 μg/kg by IP injection every day for 5 days (▪). Initially, CIDs were administered every 4 weeks. Frequency was increased to every 2 weeks at 34 weeks after transplantation. Each arrow represents a course of CID. Five control mice received injections of the carrier compound at a similar schedule (□). The 40-week injection was given to only 3 of the 5 CID-treated mice because 2 of the mice had more than 13 × 10⁹ RBCs/mL. Panel A demonstrates that the expansion of GFP-positive RBCs is dependent on the administration of CID. GFP-positive platelets are shown in panel B. Effects of CID on reticulocytes are shown in panel C. Panel D demonstrates the differences in the RBC numbers between treated and untreated animals. The absolute numbers of GFP-positive and GFP-negative RBCs are shown in panels E and F, respectively. Panel F shows the decrease in nontransduced RBCs that occurs coincident with CID administration relative to the nontreated control croup. Panel G demonstrates the separation in the 2 groups during the pretreatment phase, the period when the mice were treated every 4 weeks, and the period when the animals were treated every other week. Error bars indicate standard error of the mean, but in panel G they show 1 SD. Significance of the asterisks in panels C and D is discussed in “Results.”